Difference between revisions of "Data Annotation"
From CSWiki
| Line 3: | Line 3: | ||
Img_W Img_H xmin xmax ymin ymax | Img_W Img_H xmin xmax ymin ymax | ||
* Problems of Current Approach | * Problems of Current Approach | ||
| − | # some images are too large | + | # some images are too large to be displayed within the screen, so have bad annotation. |
# the "null" string in result should be replaced by -1 -1 -1 -1 | # the "null" string in result should be replaced by -1 -1 -1 -1 | ||
# the following plot is that, if each image has a certain number of annotations (x axis), how many images can be inferred a true bounding box. | # the following plot is that, if each image has a certain number of annotations (x axis), how many images can be inferred a true bounding box. | ||
Revision as of 19:59, 4 January 2009
- Bounding box of ImageNet images
Annotation file format: Img_W Img_H xmin xmax ymin ymax
- Problems of Current Approach
- some images are too large to be displayed within the screen, so have bad annotation.
- the "null" string in result should be replaced by -1 -1 -1 -1
- the following plot is that, if each image has a certain number of annotations (x axis), how many images can be inferred a true bounding box.
By the plot, we suggest using 6 annotations for each image to label.
